


eststo clear





* COLUMNS FROM MECHANISM EXPERIMENT 1
use "${pathdata_mechanism}/MechanismExperiment1.dta", clear
preserve

* BELIEF REGRESSION Mechanism Experiment 1
keep if product=="restaurant"


tab restaurant_version

gen lowsim1 = 0
gen lowsim2 = 0
replace lowsim1 =1 if restaurant_version == "eatery_italian"
replace lowsim2 =1 if restaurant_version == "jones_italian"



* Append the dataset with a copy of itself
expand 2, generate(newv)



* Generate the variables estimate and delay
gen belief_impact = .
gen delay = .

* Set delay=1 for the first half of the new dataset and delay=0 for the second half
replace delay = 1 if newv == 0
replace delay = 0 if newv == 1

* Set estimate=beliefs_ for the first half and estimate=p_guess_ for the second half
replace belief_impact = effect_recall if newv == 0
replace belief_impact = effect if newv == 1

gen lowsim1_x_delay = lowsim1*delay
gen lowsim2_x_delay = lowsim2*delay

tab restaurant_version

* Run the regression with clustered standard errors on prolific_pid
eststo clear
eststo reg1: regress belief_impact lowsim1 lowsim1_x_delay lowsim2 lowsim2_x_delay delay, robust cluster(prolific_pid)


restore
* RECALL REGRESSION Mechanism Experiment 1
keep if product=="restaurant"

gen lowsim1 = 0
gen lowsim2 = 0
replace lowsim1 =1 if restaurant_version == "eatery_italian"
replace lowsim2 =1 if restaurant_version == "jones_italian"


* Run the regression with clustered standard errors on prolific_pid
eststo reg3: regress correct_recall lowsim1 lowsim2, robust cluster(prolific_pid)


* COLUMNS FROM MECHANISM EXPERIMENT 2
use "${pathdata_mechanism}/MechanismExperiment2.dta", clear
preserve


* BELIEF REGRESSION Mechanism Experiment 2
keep if product=="foodtruck"

gen highsim = 0
replace highsim = 1 if treatment == "treatment"



* Append the dataset with a copy of itself
expand 2, generate(newv)



* Generate the variables estimate and delay
gen belief_impact = .
gen delay = .

* Set delay=1 for the first half of the new dataset and delay=0 for the second half
replace delay = 1 if newv == 0
replace delay = 0 if newv == 1

* Set estimate=beliefs_ for the first half and estimate=p_guess_ for the second half
replace belief_impact = effect_recall if newv == 0
replace belief_impact = effect if newv == 1

gen highsim_x_delay = highsim*delay

* Run the regression with clustered standard errors on prolific_pid
eststo reg2: regress belief_impact highsim delay highsim_x_delay, robust cluster(prolific_pid)


* RECALL REGRESSION Mechanism Experiment 2
restore

keep if product=="foodtruck"

gen highsim = 0
replace highsim = 1 if treatment == "treatment"


* Run the regression with clustered standard errors on prolific_pid
eststo reg4: regress correct_recall highsim, robust cluster(prolific_pid)
	
	
	

	
	esttab reg1 reg2 reg3 reg4 using "${pathout_mechanism}/tables/tableA10.tex", ///
		booktabs nonotes replace label nomtitles ///
		keep(lowsim1 lowsim2 highsim lowsim1_x_delay lowsim2_x_delay highsim_x_delay delay _cons) order(lowsim1 lowsim2 highsim lowsim1_x_delay lowsim2_x_delay highsim_x_delay delay) coeflabels( lowsim1 "Low Similarity 1" lowsim2 "Low Similarity 2" highsim "High Interference" lowsim1_x_delay "Delay $/times$ Low Similarity 1" lowsim2_x_delay "Delay $/times$ Low Similarity 2" highsim_x_delay "Delay $/times$ High Interference" delay "Delay"  _cons "Control Mean") ///
		se(2) b(a2) stats(x N r2, fmt(2 0 2) label("Observations" "R^2")) star(* 0.10 ** 0.05 *** 0.01) ///
		prehead("{\begin{tabular}{l*{4}{c}}\toprule\toprule&\multicolumn{4}{c}{\textit{Dependent variable:}}//[.1cm] &\multicolumn{2}{c}{Belief Impact} &  \multicolumn{2}{c}{Combined Recall} // \cmidrule(lr){2-4} \cmidrule(lr){5-5} // \textit{Sample:} &\multicolumn{1}{c}{Cue-target sim.} &\multicolumn{1}{c}{Cue-non-target sim.} &\multicolumn{1}{c}{Cue-target sim.} &\multicolumn{1}{c}{Cue-non-target sim.}//")

